Genetics 2008

Copyright � 2008 by the Genetics Society of AmericaDOI: 10.1534/genetics.107.084533

Interactions Between Stressful Environment and Gene Deletions Alleviate theExpected Average Loss of Fitness in Yeast

Lukasz Jasnos,1 Katarzyna Tomala, Dorota Paczesniak2 and Ryszard Korona3

Institute of Environmental Sciences, Jagiellonian University, 30387 Krakow, Poland

Manuscript received November 14, 2007Accepted for publication January 30, 2008

ABSTRACT

The conjecture that the deleterious effects of mutations are amplified by stress or interaction with oneanother remains unsatisfactorily tested. It is now possible to reapproach this problem systematically byusing genomic collections of mutants and applying stress-inducing conditions with a well-recognizedimpact on metabolism. We measured the maximum growth rate of single- and double-gene deletionstrains of yeast in several stress-inducing treatments, including poor nutrients, elevated temperature, highsalinity, and the addition of caffeine. The negative impact of deletions on the maximum growth rate wasrelatively smaller in stressful than in favorable conditions. In both benign and harsh environments,double-deletion strains grew on average slightly faster than expected from a multiplicative model ofinteraction between single growth effects, indicating positive epistasis for the rate of growth. Thistranslates to even higher positive epistasis for fitness defined as the number of progeny. We conclude thatthe negative impact of metabolic disturbances, regardless of whether they are of environmental or geneticorigin, is absolutely and relatively highest when growth is fastest. The effect of further damages tends to beweaker. This results in an average alleviating effect of interactions between stressful environment and genedeletions and among gene deletions.

RECENT experiments suggest that the genomic rateof spontaneous deleterious mutation is high

(Denver et al. 2004; Haag-Liautard et al. 2007). Spon-taneous mutagenesis must be countered by purgingselection—that is, the enhanced mortality or reducedfecundity of bearers of mutations—or offset by com-pensatory mutations (Silander et al. 2007). It has beenrepeatedly proposed that a harsh environment, orstress, is likely to aid selection by imposing demandsunbearable for individuals weakened by mutations.This simple and intuitively appealing assumption issupported by the results of some experiments demon-strating that the negative effects of random mutationsare higher under adverse physical conditions or severecompetition (Kondrashov and Houle 1994; Shabalina

et al. 1997; Korona 1999; Vassilieva et al. 2000;Szafraniec et al. 2001; Yang et al. 2001; Fry andHeinsohn 2002). However, not all studies confirm thisexpectation (Fry et al. 1996; Martin and Lenormand

2006) and an opposite effect has also been described(Kishony and Leibler 2003). Moreover, the reported

cases of aggravation of deleterious effects in harshenvironments are difficult to interpret. Earlier studiesoften involved organisms with large and unknownnumbers of mutations. It is thus unclear whether stressexposes more mutations or increases their averageeffects (Szafraniec et al. 2001). Finally, it is possiblethat environmental stress may promote negative (low-ering fitness) genetic interactions among deleteriousmutations. This direction of epistasis probably does notdominate under normal environmental conditions (De

Visser and Elena 2007; Kouyos et al. 2007; Jasnos andKorona 2007). However, it is unsure whether the aver-age effect of epistasis can change under stress (You andYin 2002; Kishony and Leibler 2003; Cooper et al.2005; Killick et al. 2006). In sum, the basic characteris-tics of deleterious mutations are insufficiently recognized,especially when the changeability of the environmentis taken into account. It therefore remains unclearwhether the accumulation of deleterious mutations canendanger the existence of populations (Kimura andMaruyama 1966; Crow and Kimura 1979; Schultz

and Lynch 1997) and whether the widespread occur-rence of genetic recombination and sex is an evolu-tionary response to this threat (Otto and Lenormand

2002).We chose the organism, mutations, and environments

specifically to overcome or reduce the problems typi-cally met in earlier studies. We used isogenic strains ofthe budding yeast with none, one, or two gene deletions.

1Present address: Marie Curie Research Institute, Oxted, Surrey RH80TL, United Kingdom.

2Present address: Department of Animal and Plant Sciences, University ofSheffield, Sheffield S10 2TN, United Kingdom.

3Corresponding author: Institute of Environmental Sciences, JagiellonianUniversity, Gronostajowa 7, 30387 Krakow, Poland.E-mail: [email protected]

Genetics 178: 2105–2111 (April 2008)

The use of gene deletions guaranteed that each in-troduced alteration meant the complete loss of a pro-tein in all studied environments. Another advantage ofthe laboratory yeast strains is the wealth of informationabout genetics and cell physiology provided by tradi-tional work as well as recent genomewide studies ofgene function and expression (Scherens and Goffeau

2004). Finally, in most previous studies the molecularbasis of stress reaction was poorly known. It was definedmostly in gross terms, that is, as the occurrence of addi-tional energetic costs or simply as a decrease in fitness(Martin and Lenormand 2006). In yeast, it is possibleto select environments that are known to elicit exten-sive, specific, and functionally interpretable reactions ofcellular metabolism (Bahn et al. 2007).

The laboratory environment of microbial cultures isdetermined by the applied nutrients, physical condi-tions, and additional compounds, such as drugs. In thisexperiment, one benign and four harsh environmentswere used. The two most often applied yeast nutritionalmedia are YPD, a rich resource of organic compoundsincluding amino acids and nucleotides, and SD, a syn-thetic medium containing a minimum set of vitamins,salts, and simple sources of nitrogen and organic carbon(Sherman 2002). In this study, YPD represents thebenign environment, whereas SD models nutritionalstress. Two model environmental stresses applied in thisstudy are high temperature, 37� instead of the standard30�, and a high salinity of 1 m NaCl. Faced with stress, theyeast cell reacts universally by remodeling of transcrip-tional activity, protection of existing proteins, and ad-justment of carbohydrate metabolism (Marchler et al.1993; Martinez-Pastor et al. 1996). High temperatureresults specifically in massive protein misfolding andultimately tests the cell’s ability to synthesize, maintain,and dispose of proteins with high efficiency (Mager

and De Kruijff 1995). High salinity primarily stressesthe robustness of the cell wall, cytoskeleton, and vac-uolar system (Hampsey 1997; Hohmann 2002). Thefinal environment used in this study is 0.8 mm caffeine.This compound alters the activity of PKA, Tor1, andPkc1, important regulators of cell metabolism, andinfluences the stability of chromosomes and the traf-ficking of proteins through Golgi and vacuole biogen-esis (Bianchi et al. 2001; Levin 2005; Kuranda et al.2006). It likely represents an artificial metabolic stress,one that, in contrast to deprivation of nutrients, heat, orsalinity, was not routinely encountered by the yeast in itsevolutionary past.

In this experiment we focused on genes whose dele-tion causes detectable deleterious effects under favor-able conditions. We assumed that these genes are likelyto provide ‘‘housekeeping’’ activities and therefore beimportant also in other environments. We thereforeasked whether the growth defects detectable under thebenign environment are amplified by stress. We foundthat the relative deleterious effect of single mutations

was lower under stress than in the benign environment.The joint fitness effect of two deleterious mutations wassmaller than expected from their individual effects inboth benign and stressful environments, indicatingpositive epistasis.

MATERIALS AND METHODS

Strains: We used a collection of single-gene deletionsengineered in the laboratory strain BY (Giaever et al. 2002).In Jasnos and Korona (2007), a list of 639 nonlethal genedeletions was compiled. The sole criterion used in selectingthe deletions was the annotation that the deletion generated adetectable growth effect in rich medium. The selection wasbased on results of genomewide phenotypic screens andtraditional genetic studies (Jasnos and Korona 2007). In thisstudy, not all of the originally selected gene deletions wereused (see below). The omission of some strains was decided apriori as a result of limited access to instruments, not because ofthe characteristics of the omitted strains.

Media: Rich medium was composed of standard YPD (1%yeast extract, 2% peptone, 2% glucose) at 30� and at 37�; YPDwith salinity at 1 m NaCl added at 30�; and YPD with 8 mm

caffeine added at 30�. Minimal medium was composed ofstandard SD (2% glucose, 0.6% yeast nitrogen base withoutamino acids) with uracil (20 mg/liter), histidine (20 mg/liter), methionine (20 mg/liter), leucine (100 mg/liter), andlysine (30 mg/liter) at 30�.

Growth assays: The growth curve of every progeny strain wasassayed independently twice. ½In the Jasnos and Korona

(2007) study, four replicates in YPD were obtained. Of those,only the first two of the relevant crosses were used in this studywhen comparisons with new estimates were made.�Growth wasmeasured in an automated workstation, Bioscreen C. Growthcurves were transformed with a polynomial function tocompensate for the nonlinear relation between populationdensity and the OD readings (Warringer and Blomberg

2003) and then log-normally transformed. Regions of linearrelation between the doubly transformed readings and timewere defined on the basis of the pilot studies. They were usedto calculate regression lines whose slope was equivalent tomaximum growth rate (MGR) ( Jasnos et al. 2005; Jasnos andKorona 2007). In total, 10,520 individual growth curves wereanalyzed (263 crosses, each giving four progeny strains, fiveenvironments, and two replications).

RESULTS

Growth effects of gene deletions: It was shown in anearlier study that the functional annotations of genedeletions causing growth defects followed a patterncharacteristic for the whole genome; that is, they wereroughly representative of complete cellular metabolism( Jasnos and Korona 2007). In half of the deletions, agenetic marker present in the deletion cassette (re-sistance to geneticin, or kan) was exchanged for anotherone (resistance to nourseothricin, or nat). Differentlymarked haploid strains of the opposite mating typeswere randomly matched, mated, and sporulated. Fromeach cross involving two haploid parental strains withsingle deletions, a tetrad of recombinant haploid strainswas derived: unmarked, kan marked, nat marked, and

2106 L. Jasnos et al.

both kan and nat marked. MGR of the progeny strainswas then estimated in a favorable environment of richbroth medium (YPD) at 30� ( Jasnos and Korona 2007).

In this experiment, we extended measurements ofMGR in four unfavorable environments: minimal me-dium, high saline, high temperature, and caffeine. Intotal, we report here 263 random crosses between 526(263 3 2) parental strains of which 1052 (263 3 4)progeny strains were obtained and assayed for MGR.½We assayed 263 wild-type strains (products of crosses)instead of a single ancestral strain. In this way, weensured that an average genetic background of thezero-deletion strains was exactly the same as that of thebearers of deletions, including randomly segregatingunknown polymorphism.� The new estimates werecompared to a relevant sample known from the favor-able environment. Figure 1 shows the frequency distri-butions of MGR for strains with none, one, or two

deletions in every environment. The list of genedeletions and estimates of MGR is presented in supple-mental Table 1.

To compare the growth effects of gene deletions indifferent environments, we converted the absolutevalues of maximum growth rate to relative growth rates(RGR). Individual MGRs were divided by the averageMGR of the 263 strains with no deletion (the progenywild-type strains). The standardization was done foreach environment independently. Figure 2 shows therelation between the average MGR and RGR in all testenvironments. In general, the average RGR tended tocorrelate negatively with the average MGR. In otherwords, the relative damage caused by a gene deletion(1� RGR) was high when the MGR was high. This meansthat stress tended to alleviate the negative effect of genedeletion. (Statistical comparisons between pairs of envi-ronments are reported below after an ANOVA test.)

Figure 1.—Frequency distribution of the MGR. Average values with 95% confidence limits and medians are shown. The kan-and nat-marked single deletions are pooled; therefore their samples are twice as numerous as those with none or double deletions.

Figure 2.—Averagegrowthrate effects of gene dele-tions. The average growthrate of strains containingno deletions (open bars),single deletions (shadedbars), and two deletions(solid bars) with 95% confi-dence limits are shown. Fiveenvironments were used:YPD (rich medium), heat(YPD and 37�), SD (minimalsynthetic medium), caffeine(8 mm caffeine in YPD),and saline (1 m NaCl inYPD). (A) MGR. (B) RGR.

Environmental Stress and Genetic Epistasis 2107

We then asked whether some deletions tended to bemore harmful than others when averaged across allenvironments. This analysis was restricted to strainsbearing single deletions. To be conservative, we ex-cluded deletions that produced the lethal or sublethalphenotype in any of the compared environments (seesupplemental Table 1). Results of a two-way ANOVA testcarried out for the standardized data, RGRs, are sum-marized in Table 1. The genetic factor was significant,indicating that some deletions tended to be generallymore harmful than others. The negative effect of a de-letion was modulated by the environment as evidencedby a statistically significant interaction between genedeletion and environment. The environmental factorwas also significant. An a posteriori Tukey’s test revealedthat there was no difference in the average RGR onlybetween YPD and heat. All other pairs differed signifi-cantly, P , 0.0001, yielding four clusters in an increasingorder of the average RGR: YPD together with heat,caffeine, SD, and NaCl.

A possible explanation for the observed alleviatingeffect of stress on the average effect of deleterious mu-tations could be that our sample is biased because weselected gene deletions that were previously found del-eterious in the benign environment. However, if thedeleterious effects were specific to YPD, or even antag-onistically pleiotropic, the correlation between YPD andother environments should be absent or negative.

Contrary to this expectation, the Pearson correlationcoefficient between YPD and the stressful 37�, SD,caffeine, and saline was always positive: 0.511, 0.409,0.335, and 0.328, respectively. (These statistics werehighly significant, as the sample sizes were large: 501,517, 496, and 505, respectively.) This result is compat-ible with our assumption that deleterious effects of theselected mutations are not specific for YPD but repre-sent general defects of cellular metabolism.

Epistasis for growth rate and fitness: Epistasis is ab-sent when the phenotypic effect of a gene is indepen-dent of genes in other loci. Considering deleteriousmutations, it is typically assumed that epistasis is absentwhen the proportional change in a trait caused by an in-dividual mutation is stable, irrespective of the presenceof other mutations in the same individual. This leads to amodel of multiplicative combination of growth effectscaused by gene deletions: RGRwt �RGRkn¼RGRk �RGRn,where wt, k, n, and kn denote, respectively, no deletion,kan-marked deletions, nat-marked deletions, and dou-ble deletions. Consequently, when growth rate is theconsidered phenotypic trait, epistasis can be defined asE ¼ RGRwt � RGRkn � RGRk � RGRn. Frequency dis-tributions of E in the five test environments are shownin supplemental Figure 1. Figure 3 shows that the aver-age values of E, although generally small, were positivein every environment. Averaged over all five environ-ments, the positive deviation from zero was statisticallysignificant (Figure 3).

In population genetics, the status of an individual/clone is identified not by its rate of growth but by itsfitness. Fitness, w, is the average number of progeny leftby an individual of a given genotype. In a continuouslygrowing population, an equivalent of offspring size isobtained from the expression w ¼ emt, where m is thegrowth rate and t is generation time (Crow and Kimura

1970). Thus, the growth rate is equal to loge(w) with t setto 1. Consequently, when fitness is the consideredphenotypic trait, testing for multiplicity of fitness

TABLE 1

ANOVA for RGR of single-deletion strains

Effect d.f. MS F P

Environment 4 1.5623 245.1916 ,0.0001Deletion 438 0.1692 26.5565 ,0.0001Interaction 1752 0.0398 6.2401 ,0.0001Error 2205 0.0064

MS, mean square.

Figure 3.—Average epi-static effects. The averagevalues and 95% confidencelimits of epistasis betweentwo deletions are shown.The confidence limits werecalculated using a single(average) value of epistasisfor one progeny strain.Sample sizes of the particu-lar environments, begin-ning from the left, were257, 238, 254, 235, and243 strains. The grandmean was calculated as amean of five environmentalmeans with 4 d.f. (A) Epis-tasis for growth rate, E.(B) Epistasis for fitness, e.


effects, wwtwnk¼wkwn, means testing for the additivity ofgrowth rates, RGRwt 1 RGRkn¼RGRk 1 RGRn. Epistasisfor fitness can be defined as e ¼ (RGRwt 1 RGRkn) �(RGRk 1 RGRn) ( Jasnos and Korona 2007). Fre-quency distributions of e in the five environments areshown in supplemental Figure 1. Figure 3 demonstratesthat the average e was positive in every environment andsignificantly .0 when averaged over all environments.Understandably, the values of e were higher than thoseof E. ½The predicted decline in the growth rate of thedouble mutant is 1 � RGRk � RGRn in the additivemodel, that is, steeper by RGRkRGRn than in the mul-tiplicative model, (1 � RGRk)(1 � RGRn).�

We applied two-way ANOVA to test how variation inepistasis is modulated by the environmental and genetic(pair of deletions) factors. We excluded crosses that in-cluded deletions producing the (sub)lethal phenotypein any of the compared environments. Table 2 reportsresults for e. All three determinants—environmental,genetic, and the interaction between them—were statis-tically significant. An analogous analysis for E yieldedqualitatively similar results, that is, high significance ofall three factors (not shown). Inspection of Figure 3suggests that the dependence of E and e on environ-ment followed a pattern similar to that observed for therelative effect of a single mutation (1 � RGR). That is,the epistatic effect tended to be higher in environmentsin which growth was faster, although the differenceswere not statistically significant.

DISCUSSION

The interactions of deleterious mutations with theenvironment and with one another are typically weak,making them difficult to predict, estimate, and interpret(Remold and Lenski 2004; Demuth and Wade 2006;Martin and Lenormand 2006; De Visser and Elena

2007; Kouyos et al. 2007). To circumvent these difficul-ties, we used a model system of yeast gene deletionsassayed in standardized laboratory environments. Ourrecent (Jasnos and Korona 2007) and present resultsshow a consistent pattern: the interactions of deleteri-ous mutations with a stress-inducing environment andwith one another are definitely not of the aggravating(synergistic) type. Instead, consistent bias toward allevi-ation is seen. We found that the average proportional

effect of deletion on growth (1�RGR) under heat stresswas close to that measured in the benign YPD environ-ment. In the nutritionally poorer synthetic medium,and especially in saline and caffeine, the average 1 �RGR was clearly lower than in YPD, indicating thatmutations became less deleterious in these media. Wedeliberately chose environments known to induce dif-ferent functional reactions to increase the chances ofexposing the variability of interactions. However, ourresults hint that the growth rate itself, rather than thespecificities of cellular metabolism, constituted animportant determinant of the phenotypic reaction tomutational damage. When growth was increasingly im-peded by adverse external conditions, metabolic defectscaused by mutations were decreasingly harmful.

A similar alleviation of deleterious effects of mutationby stress was previously described for Escherichia coli.Bacterial cells carried random point mutations, suppos-edly one per genome, whose negative impact on the rateof growth decreased after the addition of several anti-biotics, simulating a decline in environmental quality(Kishony and Leibler 2003). In this experiment, thealleviation of mutational harm was particularly clear-cutnot only in caffeine but also in high salt conditions. Thelatter is a typical natural stressor known to elicit cellularreactions that are highly evolved and common for manyorganisms (Hohmann 2002). It is thus remarkable thatboth prokaryotic and eukaryotic cells faced with a vari-ety of environmental stresses and mutational damagesshow a similar general pattern of gene–environmentinteractions.

In our experiment, genetic epistasis was positive onaverage under both favorable and stressful conditions.More specifically, the growth rate of double-deletionstrains was higher than predicted by the multiplicativeand especially by the additive model of how mutationaleffects accumulate. Multiplicity and additivity of growthrate effects are both simplified and arbitrarily chosenexpectations of how complex metabolic networksshould behave. Multiplicity of growth rates is a domi-nating model in theoretical and empirical studiesaiming to uncover specific functional interactions be-tween genes (Tong et al. 2004; Davierwala et al. 2005;Kelley and Ideker 2005; Collins et al. 2006; Ooi et al.2006; Boone et al. 2007; Collins et al. 2007; St. Onge

et al. 2007). The additive model of growth rate effects isspecial in the sense that it is equivalent to the multiplic-ity of fitness defined as the number of progeny (Jasnos

and Korona 2007). This definition of fitness is standardin models of population genetics, including the theo-retical work on the evolution of genetic recombinationand sexual reproduction (Otto and Lenormand 2002).Thus, while the multiplicative model is well suited as thenull hypothesis in metabolic experiments, the additivemodel is appropriate when comparisons with multicel-lular organisms or tests of population genetics theoryare to be made. In this study, the null predictions of both

TABLE 2

ANOVA for epistasis for fitness, e

Effect d.f. MS F P

Environment 4 0.1756 10.4381 ,0.0001Deletion 199 0.2650 15.7464 ,0.0001Interaction 796 0.0670 4.0391 ,0.0001Error 1000 0.0168

MS, mean square.


models were rejected in favor of positive (antagonistic,alleviating) epistasis. This result, together with the alle-viating effect of interaction between environment andsingle deletions discussed above, fits into a single gen-eral pattern. The first impairment inflicted upon themetabolism of a rapidly growing cell is relatively themost harmful irrespective of whether the insult origi-nates from a decline in quality of the external environmentor from the deregulation of the internal environment.This leads to speculation that antagonism is likely to becommon in interactions among environment, drugs,and mutations. It may be an evolved adaptation to down-regulate growth more strongly when it is rapid becausethis allows the avoidance of excessive overinvestment.Alternatively, the very construction of a complex cellularsystem makes it most sensitive when operating at itshighest rate.

Results of earlier studies often appear incompatiblewith the model suggested above. In particular, it wasfound in yeast that the effect of multiple mutations onthe growth rate was more deleterious in stressful than inbenign environments (Korona 1999; Szafraniec et al.2001). However, in these studies random mutageneiswas applied and therefore the exact number, location,and functional aspects of mutations were not known.Substitutions of amino acids, a common basis of ran-dom mutation, are ambiguous in their metabolic effect.They often result in only partial destabilization of pro-tein structures, which does not impair protein functionunder normal conditions but is critically exacerbatedunder stress (Pakula and Sauer 1989). Thus, muta-tions without phenotypic penetration under favorableconditions can emerge as new phenotypic defects understress (Hampsey 1997; Szafraniec et al. 2001). Thisdoes not occur when a known number of null mutations(here gene deletions) is considered. As the proportionof partial and null mutations is often unknown underrandom mutagenesis, the results of previous experi-ments are difficult to interpret. Another question is whyearlier studies did not unambiguously detect antago-nism among deleterious mutations. Experiments in-volving unicellular microorganisms usually showed thatmost epistatic effects were close to zero. Notably, inleading examples of such studies, fitness was defined asthe relative rate of growth and consequently the multi-plicative model of interaction was used (Elena andLenski 1997; Segre et al. 2005). Redefining the fitnessas the number of progeny and applying the additivemodel would likely shift the results toward positivevalues. Thus, we believe that the postulated generalalleviating effect of interactions between mutations andenvironment and among mutations are not contra-dicted by results of previous work.

Evolutionary biologists are concerned with the role ofepistasis in the evolution of genetic recombination. Themutational deterministic hypothesis predicts that notonly sporadic genetic recombination but also obligatory

sexual reproduction would be convincingly explained ifdeleterious mutations interacted synergistically (Feldman

et al. 1980; Kondrashov 1988; but see also Kouyos et al.2006). Empirical support for this assumption is gen-erally weak if not absent (De Visser and Elena 2007;Kouyos et al. 2007). This study provides further evi-dence against synergistic epistasis by including stressfulenvironments. Some current theories of sex and recom-bination reject the assumption of synergistic epistasisand rely on stochastic explanations. Briefly, unfavorablecombinations of new mutations and genetic back-grounds tend to be overly frequent even in relativelylarge populations. Recombination alleviates this effectand therefore its spread and maintenance is advanta-geous. However, the predicted benefit of genetic re-combination is relatively small or even very small (Hill

and Robertson 1966). It may be insufficient if the aver-age epistasis for fitness is positive and not exceedinglysmall (Keightley and Otto 2006). Here, epistasis wasrelatively high—higher than the difference between themultiplicative and additive combination of log fitness.Thus, positive (antagonistic, alleviating) epistasis is at oddswith both deterministic and stochastic models explainingthe evolution of recombination and sex (Barton 1995). Itis therefore highly desirable to test whether antagonisticepistasis applies to traits unrelated to fast growth andwhether it extends to multicellular organisms.

This work was supported by a grant from the State Committee forScientific Research of Poland (0684/P01/2006/30).

LITERATURE CITED

Bahn, Y. S., C. Xue, A. Idnurm, J. C. Rutherford, J. Heitman et al.,2007 Sensing the environment: lessons from fungi. Nat. Rev.Microbiol. 5: 57–69.

Barton, N. H., 1995 A general model for the evolution of recom-bination. Genet. Res. 65: 123–145.

Bianchi, M. M., S. Ngo, M. Vandenbol, G. Sartori, A. Morlupi

et al., 2001 Large-scale phenotypic analysis reveals identical con-tributions to cell functions of known and unknown yeast genes.Yeast 18: 1397–1412.

Boone, C., H. Bussey and B. J. Andrews, 2007 Exploring geneticinteractions and networks with yeast. Nat. Rev. Genet. 8: 437–449.

Collins, S. R., M. Schuldiner, N. J. Krogan and J. S. Weissman,2006 A strategy for extracting and analyzing large-scale quanti-tative epistatic interaction data. Genome Biol. 7: R63.

Collins, S. R., K. M. Miller, N. L. Maas, A. Roguev, J. Fillingham

et al., 2007 Functional dissection of protein complexes involvedin yeast chromosome biology using a genetic interaction map.Nature 446: 806–810.

Cooper, T. F., R. E. Lenski and S. F. Elena, 2005 Parasites and mu-tational load: an experimental test of a pluralistic theory for theevolution of sex. Proc. Biol. Sci. 272: 311–317.

Crow, J. F., and M. Kimura, 1970 An Introduction to Population Genet-ics Theory. Harper & Row, New York.

Crow, J. F., and M. Kimura, 1979 Efficiency of truncation selection.Proc. Natl. Acad. Sci. USA 76: 396–399.

Davierwala, A. P., J. Haynes, Z. Li, R. L. Brost, M. D. Robinson

et al., 2005 The synthetic genetic interaction spectrum of essen-tial genes. Nat. Genet. 37: 1147–1152.

Demuth, J. P., and M. J. Wade, 2006 Experimental methods for mea-suring gene interactions. Annu. Rev. Ecol. Evol. Sys. 37: 289–316.

Denver, D. R., K. Morris, M. Lynch and W. K. Thomas, 2004 Highmutation rate and predominance of insertions in the C. elegansnuclear genome. Nature 430: 679–682.


de Visser, J. A., and S. F. Elena, 2007 The evolution of sex: empir-ical insights into the roles of epistasis and drift. Nat. Rev. Genet.8: 139–149.

Elena, S. F., and R. E. Lenski, 1997 Test of synergistic interactionsamong deleterious mutations in bacteria. Nature 390: 395–398.

Feldman, M. W., F. B. Christiansen and L. D. Brooks, 1980 Evo-lution of recombination in a constant environment. Proc. Natl.Acad. Sci. USA 77: 4838–4841.

Fry, J. D., and S. L. Heinsohn, 2002 Environment dependence ofmutational parameters for viability in Drosophila melanogaster.Genetics 161: 1155–1167.

Fry, J. D., S. L. Heinsohn and T. F. C. Mackay, 1996 The contribu-tion of new mutations to genotype-environment interaction forfitness in Drosophila melanogaster. Evolution 50: 2316–2327.

Giaever, G., A. M. Chu, L. Ni, C. Connelly, L. Riles et al.,2002 Functional profiling of the Saccharomyces cerevisiae ge-nome. Nature 418: 387–391.

Haag-Liautard, C., M. Dorris, X. Maside, S. Macaskill, D. L. Halligan

et al., 2007 Direct estimation of per nucleotide and genomic dele-terious mutation rates in Drosophila. Nature 445: 82–85.

Hampsey, M. A., 1997 Review of phenotypes in Saccharomyces cerevi-siae. Yeast 13: 1099–1133.

Hill, W. G., and A. Robertson, 1966 The effects of linkage on thelimits to artificial selection. Genet. Res. 8: 269–294.

Hohmann, S., 2002 Osmotic stress signaling and osmoadaptation inyeasts. Microbiol. Mol. Biol. Rev. 66: 300–372.

Jasnos, L., and R. Korona, 2007 Epistatic buffering of fitness loss inyeast double deletion strains. Nat. Genet. 39: 550–554.

Jasnos, L., P. Sliwa and R. Korona, 2005 Resolution and repeat-ability of phenotypic assays by automated growth curve analysisin yeast and bacteria. Anal. Biochem. 344: 138–140.

Keightley, P. D., and S. P. Otto, 2006 Interference among delete-rious mutations favours sex and recombination in finite popula-tions. Nature 443: 89–92.

Kelley, R., and T. Ideker, 2005 Systematic interpretation of geneticinteractions using protein networks. Nat. Biotechnol. 23: 561–566.

Killick, S. C., A. M. Carlsson, S. A. West and T. J. Little,2006 Testing the pluralist approach to sex: the influence of en-vironment on synergistic interactions between mutation load andparasitism in Daphnia magna. J. Evol. Biol. 19: 1603–1611.

Kimura, M., and T. Maruyama, 1966 The mutational load with ep-istatic gene interactions in fitness. Genetics 54: 1337–1351.

Kishony, R., and S. Leibler, 2003 Environmental stresses can alle-viate the average deleterious effect of mutations. J. Biol. 2: 14.

Kondrashov, A. S., 1988 Deleterious mutations and the evolutionof sexual reproduction. Nature 336: 435–440.

Kondrashov, A. S., and D. Houle, 1994 Genotype-environment in-teractions and the estimation of the genomic mutation rate inDrosophila melanogaster. Proc. Biol. Sci. 258: 221–227.

Korona, R., 1999 Genetic load of the yeast Saccharomyces cerevisiaeunder diverse environmental conditions. Evolution 53: 1966–1971.

Kouyos, R. D., S. P. Otto and S. Bonhoeffer, 2006 Effect of vary-ing epistasis on the evolution of recombination. Genetics 173:589–597.

Kouyos, R. D., O. K. Silander and S. Bonhoeffer, 2007 Epistasisbetween deleterious mutations and the evolution of recombina-tion. Trends Ecol. Evol. 22: 308–315.

Kuranda, K., V. Leberre, S. Sokol, G. Palamarczyk and J. Francois,2006 Investigating the caffeine effects in the yeast Saccharomycescerevisiae brings new insights into the connection between TOR,PKC and Ras/cAMP signalling pathways. Mol. Microbiol. 61:1147–1166.

Levin, D. E., 2005 Cell wall integrity signaling in Saccharomyces cere-visiae. Microbiol. Mol. Biol. Rev. 69: 262–291.

Mager, W. H., and A. J. De Kruijff, 1995 Stress-induced transcrip-tional activation. Microbiol. Rev. 59: 506–531.

Marchler, G., C. Schuller, G. Adam and H. Ruis, 1993 A Saccha-romyces cerevisiae UAS element controlled by protein kinase A ac-tivates transcription in response to a variety of stress conditions.EMBO J. 12: 1997–2003.

Martin, G., and T. Lenormand, 2006 The fitness effect of muta-tions across environments: a survey in light of fitness landscapemodels. Evolution 60: 2413–2427.

Martinez-Pastor, M. T., G. Marchler, C. Schuller, A. Marchler-Bauer, H. Ruis et al., 1996 The Saccharomyces cerevisiae zinc fin-ger proteins Msn2p and Msn4p are required for transcriptionalinduction through the stress response element (STRE). EMBO J.15: 2227–2235.

Ooi, S. L., X. Pan, B. D. Peyser, P. Ye, P. B. Meluh et al., 2006 Globalsynthetic-lethality analysis and yeast functional profiling. TrendsGenet. 22: 56–63.

Otto, S. P., and T. Lenormand, 2002 Resolving the paradox of sexand recombination. Nat. Rev. Genet. 3: 252–261.

Pakula, A. A., and R. T. Sauer, 1989 Genetic analysis of protein sta-bility and function. Annu. Rev. Genet. 23: 289–310.

Remold, S. K., and R. E. Lenski, 2004 Pervasive joint influence ofepistasis and plasticity on mutational effects in Escherichia coli.Nat. Genet. 36: 423–426.

Scherens, B., and A. Goffeau, 2004 The uses of genome-wide yeastmutant collections. Genome Biol. 5: 229.

Schultz, S. T., and M. Lynch, 1997 Mutation and extinction: therole of variable mutational effects, synergistic epistasis, benefi-cial mutations, and degree of outcrossing. Evolution 51: 1363–1371.

Segre, D., A. Deluna, G. M. Church and R. Kishony, 2005 Mod-ular epistasis in yeast metabolism. Nat. Genet. 37: 77–83.

Shabalina, S. A., L. Y. Yampolsky and A. S. Kondrashov,1997 Rapid decline of fitness in panmictic populations of Dro-sophila melanogaster maintained under relaxed natural selection.Proc. Natl. Acad. Sci. USA 94: 13034–13039.

Sherman, F., 2002 Getting started with yeast. Methods Enzymol.350: 3–41.

Silander, O. K., O. Tenaillon and L. Chao, 2007 Understandingthe evolutionary fate of finite populations: the dynamics of mu-tational effects. PLoS Biol. 5: e94.

St. Onge, R. P., R. Mani, J. Oh, M. Proctor, E. Fung et al.,2007 Systematic pathway analysis using high-resolution fitnessprofiling of combinatorial gene deletions. Nat. Genet. 39: 199–206.

Szafraniec, K., R. H. Borts and R. Korona, 2001 Environmentalstress and mutational load in diploid strains of the yeast Saccha-romyces cerevisiae. Proc. Natl. Acad. Sci. USA 98: 1107–1112.

Tong, A. H. Y., G. Lesage, G. D. Bader, H. M. Ding, H. Xu et al.,2004 Global mapping of the yeast genetic interaction network.Science 303: 808–813.

Vassilieva, L. L., A. M. Hook and M. Lynch, 2000 The fitness ef-fects of spontaneous mutations in Caenorhabditis elegans. Evolu-tion 54: 1234–1246.

Warringer, J., and A. Blomberg, 2003 Automated screening in en-vironmental arrays allows analysis of quantitative phenotypic pro-files in Saccharomyces cerevisiae. Yeast 20: 53–67.

Yang, H. P., A.Y. Tanikawa, W. A. Van Voorhies, J. C. Silva and A. S.Kondrashov, 2001 Whole-genome effects of ethyl methanesul-fonate-induced mutation on nine quantitative traits in outbredDrosophila melanogaster. Genetics 157: 1257–1265.

You, L., and J. Yin, 2002 Dependence of epistasis on environmentand mutation severity as revealed by in silico mutagenesis ofphage t7. Genetics 160: 1273–1281.

Communicating editor: D. Charlesworth


Documents

Genetics 2008